Statistical method for analysis of membrane filter samples of airborne asbestos

Assessments of airborne fibre concentrations are customarily based on microscope evaluation of representative fields of view on membrane filter samples. Comparisons of concentrations estimated in this way should take account of random variations in the distribution of fibres or particles on the filter and of the numerical scaling necessary to convert counts to concentrations.

These factors can be taken into consideration by a general method of analysis, based on the principles of maximum-likelihood estimation, which is a generalization of the standard techniques of analysis of variance and multiple regression.

An example is given of an application of this method to the analysis of an experiment in which 16 samples were counted by five laboratories. The aims of the experiment were to examine differences in counting levels between laboratories and to quantify differences between two mounting methods used to prepare the microscope slides.

Some special cases and variations are discussed and computational details are provided.